[automatically translated] Recent climate change and more frequent floods make it particularly timely use of early warning systems to facilitate the evacuation of the population and the safety of property. These systems are based partly on short-term forecasting of rainfall, partly on estimating the rainfall-runoff and flood wave propagation process. In this memory the 2D model diffusive MAST to spatial and temporal progress already proposed by some of the authors is further improved for a more rapid solution and a better adaptability to the use of parallel computing. It is introduced a new semi-analytical solution of the convective component of the problem, with a remarkable reduction of the corresponding computational effort. The diffusive component is discretized in space and time with a classical procedure of the finished elements which results, at each time step, in a well-conditioned linear system that can be solved effectively even with the help of parallel computing. The model was implemented in the early warning system of Genna River, in Umbria. The hydrographs full planned for the short term on the basis of the measured rains which are used to condition the upstream contour of the model, whose domain covers the potentially floodable surface for the entire stretch considered to hydraulic risk.
|Number of pages||0|
|Publication status||Published - 2010|